bond-selective dissociation of polyatomic cations in mid-infrared strong fields · 2014-01-23 ·...

Bond-Selective Dissociation of Polyatomic Cations in Mid-InfraredStrong FieldsSuk Kyoung Lee, H. Bernhard Schlegel, and Wen Li*

Department of Chemistry, Wayne State University, 5101 Cass Avenue, Detroit, Michigan 48202, United States

ABSTRACT: Strong field-induced dissociation by intense mid-infrared pulses wasinvestigated in bromofluoroform monocation (CF3Br

+) and iodobenzene dication(C6H5I

2+) using ab initio molecular dynamics calculations. In both systems, bond-selective dissociation was achieved using appropriate laser polarizations andwavelengths. For CF3Br

+, energetically disfavored fluorine elimination was stronglyenhanced at wavelengths of 7 to 8 μm with polarization along a C−F bond. This is theresult of two effects: the deposition of high enough kinetic energy into the molecule bythe laser field and the near-resonant excitation of the C−F stretching mode. At shorterand off-resonant wavelengths, bromine elimination becomes significant due to rapidintramolecular vibrational energy redistribution (IVR). For C6H5I

2+, the branching ratiosfor the dissociation of the ortho-, meta-, and para-hydrogens can be controlled simply bychanging the laser polarization. These results show the general applicability of bondselective dissociation of cations by intense mid-infrared laser fields.

1. INTRODUCTIONUnimolecular reactions are essential chemical processes andhave been extensively studied to understand the moleculardynamics in detail and to find ways for controlling specificreactions. Highly energetic molecules can be readily preparedwith strong field lasers and readily undergo unimoleculardissociation. Such dissociations often require large nuclearrearrangements with concerted motion corresponding to thereaction coordinate and can occur with or without barriers toform products. Therefore, the deposition of energy into aparticular vibrational mode closely coupled to the reactionpathway could drive a molecule along a path toward the desiredproducts.1 As the most intuitive control method of chemicalreactions, bond-selective decomposition has gained muchattention over the years. It was hoped that the excitation of astretching vibrational mode using a tunable laser could achievebond-selective chemistry. However, it was realized that theinitial energy deposited in a particular bond could quickly flowto other vibrational modes on the time scale of the reaction. Inparticular, infrared multiphoton dissociation (IRMPD) of apolyatomic system is subject to rapid intramolecular vibra-tional-energy redistribution (IVR)2,3 and thus generally doesnot achieve mode selectivity. This is the reason why IRMPDcan be described well by statistical methods such as RRKMtheory. For example, Bloembergen et al. addressed how IVRcan affect the reactivity of a polyatomic system such as CF3Brwhen tuning the laser frequency close to the C−F stretchingvibration.4 IVR can lead to an ergodic ensemble within a fewpicoseconds, resulting in the breaking of the weakest bond, thatis, the C−Br bond. Thus, the products of IRMPD with IVR arethe same as a thermal reaction. The key for achieving bond-selective chemistry is to suppress IVR. With the advent offemtosecond lasers, defeating IVR becomes possible in timedomain. Zewail and coworkers5 have shown nonstatistical

behavior of fragmentation due to ultrafast electronic activation.However, this was accomplished through the limited vibronicstates accessible from the Franck−Condon region. Kompa andcoworkers6 reported that direct excitation of the vibrationmodes coupled to a reaction coordinate by a mid-IR laser pulsedrives selective bond breakage on the ultrashort time scale,preventing IVR. A different approach, termed coherent control,has been used to direct chemical reactions using shaped laserpulses and feedback mechanisms.7−10 Adiabatic passagetechniques, such as Raman chirped adiabatic passage(RCAP)11 and stimulated Raman adiabatic passage (STIR-AP),12 have been studied for controlling the population ofexcited states and thus chemical reactions. Because theefficiency of population transfer depends on the adiabaticityconditions for chirping, their time scales are limited toapproximately picoseconds, hence defeated by IVR. Combiningultrashort IR and UV laser pulses, the control of bond breakinghas also been proposed, in which the vibrationally excited wavepackets prepared by an IR pulse are transferred to the excitedelectronic state by a well-timed UV pulse to yield a targetproduct.13,14 Without exploiting the coherent properties andthe ultrafast time resolution of femtosecond laser pulses,vibrational excitation in the ground state has been shown toinfluence final products resulting from the excitation to higherelectronically states, followed by dissociation,15,16 even thoughsuch an excitation scheme does not warrant an ultrashortreaction time.A new mechanism for bond-specific dissociation was

proposed using intense, ultrashort mid-infrared laser pulses inour previous study.17 This approach differs from previous

Received: April 18, 2013Revised: August 12, 2013Published: October 7, 2013

Article

pubs.acs.org/JPCA

© 2013 American Chemical Society 11202 dx.doi.org/10.1021/jp4038649 | J. Phys. Chem. A 2013, 117, 11202−11209

pubs.acs.org/JPCA

experiments with mid-IR lasers mainly through the use ofshorter pulses and higher laser intensities − about five orders ofmagnitude higher than those of conventional IRMPD. Such ahigh laser intensity leads to significant charge redistributionamong constituent atoms and thus alters the molecularpotential. Furthermore, because of the long wavelength,atoms could gain a large amount of kinetic energy from thelaser field, promoting the chosen reaction on a few hundreds offemtoseconds time scale. The cycle-averaged kinetic energy(i.e., the ponderomotive energy, Upm = q2Emax

2 /(4 μω2), where qis the charge, Emax is the maximum electric field strength, μ isthe reduced mass, and ω is the laser frequency) scalesquadratically with the wavelength, and this becomes significantwith mid-IR laser pulses (a few electronvolts for light atomssuch as hydrogen). The large kinetic energy forces thedissociation to complete in a very short time and thus preventsIVR from playing a role in the dissociation dynamics. Anangular-dependent dissociation yield was observed in formylchloride cation: the dissociation of a specific bond isdramatically enhanced when the laser polarization is alignedwith the bond.17 This implies that breaking a chosen bond canbe accomplished simply by directing the polarization of intensemid-IR pulses parallel to the bond and thus promptlydepositing a large amount of energy in the bond.In this study, we demonstrate how generally applicable this

new bond-selective approach is. For two large target systems(CF3Br

+ and C6H5I2+), we observed promising bond selectivity,

in which energetically disfavored reaction channels dissociatepreferentially. We further discuss the benefits and challenges ofapplying our approach for the bond-selective chemistry.

2. METHODAb initio molecular dynamics (AIMD) was employed tosimulate strong-field-induced dissociation. Calculations werecarried out with the development version of the Gaussian seriesof programs18 using the B3LYP/6-311G(d,p) level of theory.For iodine atom, the Stuttgart effective core potential (SDDECP) was used for the core electrons. The classical trajectorieswere computed on the Born−Oppenheimer surface with therecently modified Hessian-based predictor-corrector (HPC)method, which enabled us to study the molecular dynamics ofpolyatomic systems in strong fields with improved timeefficiency.19 The calculations were carried out on the adiabaticground electronic state of the two systems (doublet for CF3Br

+

and triplet for C6H5I2+). The laser−molecule interaction was

modeled using an oscillating electric dipole field. Field-inducedtransitions between surfaces were not included. Trajectorieswere integrated with a time step of 0.5 fs, and the Hessian wasupdated for 10 steps before being recalculated analytically.The electric field was applied along specific directions in the

molecular frames of CF3Br+ and C6H5I

2+, as shown in Figure1A. The maximum field strengths employed in this study are0.07 and 0.09 au, corresponding the intensities of (1.7 and 2.8)× 1014 W/cm2, respectively. A five-cycle pulse in the mid-IRrange (λ = 6−12 μm) with linear polarization and a Gaussianenvelope was used in the simulations; the time evolution of thepulse is shown in Figure 1B. The initial molecular geometrieswere given zero-point vibrational energy with randomlyselected phases and no rotational energy. For each differentfield orientation and wavelength, a total of 100 trajectories wereintegrated for up to 0.75 ps. Trajectories were terminated whenthe distance of atoms in the fragments was over 10 bohrs (5.3Å).

In a strong field with a laser intensity of 1014 W/cm2, neutralmolecules are ionized rapidly. Therefore, we chose cations ordications as target systems to avoid the ionization process fromcomplicating the dissociation process. The ionization proba-bilities for cations or dications are much smaller than those forneutrals due to considerably higher ionization potentials (IPs).Simulations were carried out for the mono cation ofbromofluoroform; the triplet dication was chosen foriodobenzene. For the laser pulse selected in this study, weestimated the ionization probability by integrating theionization rate over the laser pulse. The ionization rate wascalculated by a modified ADK (Ammosov−Delone−Krainov)model,20 in which the ionization rate is associated with the IP.The IPs of the systems (IP = 19.7 and 19.6 eV for CF3Br

+ andC6H5I

2+, respectively), were computed at the CCSD(T)/6-311+G(2df,2pd) level of theory. The resulting ionizationprobabilities of both systems are ∼3% or less with themaximum field strength of 0.09 au, which can be neglectedwhen compared with the almost 100% dissociation proba-bilities. However, the estimated ionization probabilities wereobtained from IPs calculated at the equilibrium geometries andthus are lower limits. We should account for possible enhancedionization for deformed geometries during the laser pulse. Withthe short, intense pulses used in the present study, most of thedissociations occur after the pulse, thereby reducing thepossibility of further ionization, as will be discussed later.

3. RESULTBromofluoroform Cation (CF3Br

+). The photodissocia-tion of neutral bromofluoroform is a classical example in whichthe mode selectivity using IRMPD has failed due to

Figure 1. (A) Bromofluoroform monocation (CF3Br+) and

iodobenzene dication (C6H5I2+) structures and the field directions

defined for classical trajectory simulations. (B) Temporal evolution ofthe laser electric field with the laser intensity of Imax = 1.7 × 1014

W/cm2 (corresponding field strength is 0.07 au) at 7 μm.

The Journal of Physical Chemistry A Article

dx.doi.org/10.1021/jp4038649 | J. Phys. Chem. A 2013, 117, 11202−1120911203

anharmonic coupling.4 We show that strong field dissociationby mid-IR laser can achieve this in its cation form. Thedissociation energy for bromine elimination channel (0.54 eV,calculated for CF3Br

+ → CF3+ + Br at the B3LYP/6-311G(d,p)

level of theory) is about one-quarter of the energy for fluorineelimination (2.12 eV), and thus bromine dissociation shoulddominate according to statistical predictions. The calculationfor randomly aligned CF3Br

+ was first performed in a weakcontinuous-wave mid-IR laser field (Imax = ∼2 × 1012 W/cm2 at7 μm). The results showed that 21 out of 100 trajectoriesyielded C−F bond dissociation and 79 yielded C−Br bonddissociation. Accounting for the fact that there are three C−Fbonds and one C−Br bond, this gives a C−Br to C−Fbranching ratio of 11:1. This is due to IVR among the coupledvibrational modes (most of the trajectories (93%) are completewithin 5 ps). It can be expected that at an even lower intensitythe branching ratio for C−Br bond dissociation would be evenlarger. In contrast, at higher laser intensities and fixed nuclearorientation, the results are dramatically different.We investigated the normalized yields under different

conditions by varying the intensity, the wavelength and thelaser polarization along a chosen bond. (See Figure 1A.) In thisstudy, the normalized yield was calculated by dividing thenumber of trajectories leading to the specific channel by thetotal number of trajectories. The wavelength dependence of thenormalized yields for polarization along a C−F bond issummarized in Figure 2A for two different field strengths. Theresults clearly show that F elimination is dominant over therange of wavelengths, exhibiting nonstatistical behavior andbond selectivity. The yield of F product has a broad maximumat 7 to 8 μm. The peak becomes somewhat narrower when themaximum field strength is lowered to 0.07 au (and evennarrower at 0.06 au (1.3 × 1014 W/cm2), not shown in Figure2A). This enhancement of the F yield can be attributed toresonant excitation of the C−F antisymmetric stretchingvibrational modes (6.9 μm; all vibrational normal modes arelisted in Table 1). With the high laser intensity used in ourcalculation, it can be expected that the molecular potential islargely suppressed and thus vibration frequencies should bered-shifted, which is consistent with the peak at longerwavelength. Nevertheless, we still observe resonance behaviorat such a high intensity. This resonance effect adds anadditional control “knob” to our approach using differentmid-IR wavelengths. Recently, a similar resonance behavior hasbeen observed in a numerical solution of time-dependentSchrodinger equation (TDSE) of H2

+ in strong IR fields, inwhich the dissociation probability increases around thewavelength corresponding to the vibration frequency (alsored-shifted).21 The original idea of IRMPD was that resonantexcitation of a specific vibrational mode should correlate withthe selected dissociation pathway. The present results indicatethat excitation of a particular vibrational mode does play a rolein the dynamics of strong-field-induced dissociation, eventhough the resonance is much broader due to the distortion ofthe molecular potential by the intense laser field.The Br elimination occurs generally as a minor channel when

the laser polarization is along the C−F bond but is over 20% at6 μm with the maximum field strength of 0.07 au. At this off-resonance wavelength, Br elimination could be competing withF elimination due to a rapid IVR. It should also be noted that asubstantial number (42%) of trajectories do not dissociatewithin 750 fs. (We refer to these as unfinished trajectorieshere.) If given enough time, these are likely to result in breaking

of the weak C−Br bond after IVR. For example, when weextended the calculation time of a few unfinished trajectories to10 ps, 70% of them led to Br elimination but none to Felimination, with the rest still not finished within 10 ps. Thetotal energy gained (ΔE = ΔEkin + ΔEpot) can indicate whetherdissociation can eventually happen. The blue curve in Figure 3shows that many of the unfinished trajectories have enoughenergy for Br dissociation at 6 μm and the maximum fieldstrength of 0.07 au. On the basis of this information, weconcluded that Br is produced mainly via IVR at 6 μm,especially at lower laser intensities.

Figure 2. Normalized dissociation yields from bromofluoroformmonocations as a function of wavelength: (A) laser polarization alonga C−F bond and (B) laser polarization along the C−Br bond. Bluelines represent F eliminations and red lines represent Br eliminations.Closed symbols are for a maximum field strength of 0.07 au; opensymbols are for a maximum field strength of 0.09 au. Each simulationconsisted of 100 trajectories. The dotted lines in panel A correspondto the normalized yield of the F aligned with the laser field. Purplediamonds in panel B correspond to the normalized yield of multibodydissociation with field strengths 0.07 (closed symbols) and 0.09 au(open symbols).

Table 1. Calculated Vibrational Normal Modes of CF3Br+

under Field-Free Condition at the B3LYP/6-311G(d,p)Level of Theory

wavenumber (cm−1) wavelength (μm)

120 83.3 C−Br stretching139 72.1 bending210 47.6 bending556 18.0 bending559 17.9 bending677 14.8 bending960 10.4 C−F sym. stretching1449 6.9 C−F antisym. stretching1452 6.9 C−F antisym. stretching



Figure 2B shows the normalized yields with respect to thedriving laser wavelength when the polarization lies along theC−Br bond. The total dissociation probability increases as thewavelength becomes longer and the field strength becomeshigher, which is consistent with the fact that the ponderomotiveenergy is quadratically proportional to the wavelength andlinearly proportional to the laser intensity. It is also worthnoting that dissociation occurs less when the polarization isalong the C−Br bond than when the polarization is along a C−F bond. This is because the ponderomotive potential isinversely proportional to the mass of a constituent atom: theheavier bromine ions gain less kinetic energy from the laserfield than the lighter F ions. Figure 3 shows that at 6 μm, with amaximum field strength of 0.07 au and the field aligned with theC−Br bond, most of the unfinished trajectories gain less energythan needed for Br elimination (i.e., these trajectories will notdissociate even with a longer simulation time). For laserpolarization along the C−Br bond, we observed reasonablebond selectivity at 8 μm with the higher laser intensity, while no

bond selectivity was seen at lower intensities with any of thewavelengths considered. These results indicate that higher fieldstrengths are required to selectively break the C−Br bond.However, the higher field strength can cause significantionization before dissociation, which suggests that our methodmay not be efficient for Br dissociation. However, because C−Br bond is the weaker bond, conventional thermal activation orIRMPD should work efficiently.An interesting result is observed at the longer wavelength (10

μm) for both field directions (along C−Br or C−F bonds):when the polarization is along the C−Br bond, the dominantevent is multibody dissociation, in which more than twofragments are produced (see Figure 2B; we considered thatbond fission happens when C−F bond lengths reach 3.8 Å(three times the equilibrium distance of C−F) and C−Br bondlengths reach 4.7 Å (two times the equilibrium distance of C−Br), respectively); when the polarization is along a chosen C−Fbond (Figure 2A), the dissociation reaches 15−20% for C−Fbonds not aligned with the laser field and over 10% for C−Brbond, while at shorter wavelengths the chosen C−F bondexclusively dissociates. However, the major dynamical processesleading to these two observations are quite different and wediscuss them as follows.If the C−Br bond lies along the laser polarization, two effects

could contribute multibody dissociation: (1) a higher ponder-omotive energy and (2) a near-resonance with the C−Fsymmetric stretching (10.4 μm). Because of the largeponderomotive energy at 10 μm, even constituent atoms thatare not aligned with the field can gain enough energy todissociate, especially the lighter atoms. Furthermore, polar-ization along the C−Br bond overlaps with the dipole momentassociated with the C−F symmetric stretching mode, enhancingthe resonant excitation of this vibrational mode. We furtherinvestigate these two aspects by inspecting the time-dependentcharge evolutions and angular distribution of product, shown inFigure 4. Indeed, the charge distribution along the C−Br bondchanges the most, and fluorine charges fluctuate much more at10 μm than at 7 μm, implying little or no energy gain from thelaser field at the shorter wavelength. If fluorine atoms dissociate

Figure 3. Total energy (ΔE = ΔEkin + ΔEpot) gained from the laserfield (maximum field strength of 0.07 au and wavelength of 6 μm, 100trajectories). Red squares and blue circles correspond to Br eliminationand unfinished trajectories, respectively, when the polarization is alonga C−F bond. Green diamonds correspond to unfinished trajectorieswhen the polarization is along the C−Br bond. The vertical dashedline indicates the threshold dissociation energy of Br elimination.

Figure 4. Charge evolution in bromofluoroform ions and the angular distributions of Br and F products when the field is directed along the C−Brbond with the maximum field strength of 0.09 au. (A) Temporal evolution of the Mulliken charge on C (black), Br (red), and F (blue, green, andpurple). The top and bottom panels represent charge evolution at 7 and 10 μm, respectively. (B) Angular distributions of recoiled Br (red) and F(blue) atoms at 10 μm (100 trajectories); θ is the angle between the recoil direction and the polarization axis (C−Br bond). The fluorine angulardistribution was obtained for all three fluorine atoms.



directly by the energy deposited from the laser field, then weshould see fluorine atoms scattered along the polarization axisbecause they preferentially dissociate when they are closelyaligned with the laser field. However, fluorine has a broadangular distribution, as seen in Figure 4B, and no fluorine atomis observed within 45° of the laser polarization. By contrast, theangular distribution of bromine atoms is very sharp, which isthe character of the direct dissociation. This suggests that adirect dissociation mechanism for C−F dissociation due to alarge ponderomotive potential can be ruled out. Additionalcalculations at 12 μm (not shown) yield 100% Br dissociationand no multibody dissociation. Therefore, we conclude that C−F bond multibody dissociations observed at 10 μm are mainlydue to resonant excitation of the symmetric C−F stretch.When the polarization is along a chosen C−F bond, most of

the energy deposited into the system results from the largecharge oscillation in the C−F bond aligned with the field. Thisenergy can flow to other vibrational modes via IVR, leading toindirect dissociation of the remaining bonds in addition to thedirect dissociation of the chosen C−F bond. This is similar tothe results from a 6 μm driving laser with the polarization alonga C−F bond. The main difference is the production of F atomnot aligned with the field at 10 μm, but such dissociation isabsent at 6 μm. This is likely due to a higher ponderomotivepotential at a longer wavelength. As previously mentioned, thefluorines not aligned with the laser field could gain energy atthis wavelength; bromine would gain significantly less energybecause of its heavier mass. Even though the energy gainedfrom the field is not enough to dissociate the C−F bondsdirectly because they are not in resonance, the energy is mainlylocalized in C−F bonds and the molecules clearly show adifferent branching ratio from those at 6 μm. The multibodydissociation that is dominant with the polarization along theC−Br bond was not observed. This is not surprising becausethe vibrational transition moment (symmetric stretching) isnearly perpendicular to this field direction.It is important to note that at 8 and 10 μm with the

maximum field strength of 0.09 au, some of the fluorine yielddissociates within the last laser cycle. For shorter wavelengths at0.09 au and for all wavelengths studied at a field strength of0.07 au, the dissociation of CF3Br

+ occurs after the laser pulse isoff. In our previous study we referred to this as field-freedissociation.17 Because of the high field strength, a largeamount of kinetic energy is quickly deposited into specificvibrational modes during the pulse without significantlychanging the nuclear geometry, and this is followed by largeamplitude stretching of the bond after the pulse. In thissituation, we can safely rule out possible increased ionizationdue to charge-resonance-enhanced ionization (CREI), which

becomes significant only when the bond is substantiallyelongated. The molecule remains near the equilibriumgeometry during the pulse, and as previously mentioned, theionization probability at this geometry is small compared withdissociation.

Iodobenzene Dication (C6H5I2+). As we have shown so

far, our method depends on the capability of varying the laserpolarization in the molecular frame; however, it is not trivial toachieve molecular alignment or orientation in the gas phase,which itself is still an active research topic. However,iodobenzene is one of the few systems in which a very highdegree of spatial alignment has been achieved experimentally.Laser-induced alignment of this molecule has been obtained byadiabatic excitation22 and reported to reach ⟨cos2θ⟩ = 0.97 withsampled molecules in low-rotational temperature.23 Hence, thismolecule is a good experimental candidate for demonstratingbond-selective dissociation by strong mid-IR fields. In thisstudy, the dication was selected because the monocation has alow IP and can be easily ionized further with the laser intensityused here. The ground electronic state of iodobenzene dicationis a triplet state. The dissociation energies are 0.7 eV for I+

elimination and 4.4 eV for H+ elimination (averaged over theortho, meta, and para pathways, computed for C6H5I

2+ at theB3LYP/6-311G(d,p) level of theory).Compared with CF3Br

+ and the previous study of ClCHO+17,the threshold dissociation energies of the desired channels(selective H+ elimination) are quite high, and thus we lookedinto the dynamics at long wavelengths to take advantage of thehigh ponderomotive energy. The normalized yields calculatedat 10 μm are listed in Table 2. With a maximum field strengthof 0.09 au and polarization along the C−I bond, para-hydrogenelimination occurs exclusively, whereas only ortho- and meta-hydrogen eliminations take place when the polarization isperpendicular to the C−I bond. Because of the highpolarizability of the iodine atom, the field-induced chargeoscillation is the strongest when the polarization is along theC−I bond. Furthermore, because hydrogen is the lightest atom,it gains much more kinetic energy from the laser field thaniodine atoms, and thus selective dissociation can be achieved.The calculation results show that a high degree of control of thenormalized yields of ortho and meta versus para hydrogendissociation can be achieved by manipulating the laserpolarization.With a maximum field strength of 0.09 au, however, the

competition from ionization should be carefully considered.Even though the estimated ionization rate at the equilibriumgeometry is low for the selected laser pulses, strong fieldionization can be amplified by CREI as bond lengths arestretched. Unlike the case of CF3Br

+, most of the iodobenzene

Table 2. Normalized Yields for Dissociation Channels, Dissociation Probabilities and Average H+ Dissociation Times ofIodobenzene Dication (C6H5I

2+) for Different Laser Pulse Parameters at a Wavelength of 10 μm

channel

polarization I+ ortho-H+ meta-H+ para-H+ Pd avg. time for H+ dissociation (fs)c

Field Strength of 0.09 auparallel to C−Ia 0 0 0 1 1 84perpendicular to C−Ia 0 0.49 0.35 0 0.85 117Field Strength of 0.07 auparallel to C−Ia 0.09 0 0 0.84 0.93 114parallel to C−Ib 0.05 0 0.01 0.57 0.63 104

aFive-cycle Gaussian pulse with a phase of 0°, 100 trajectories. bFour-cycle Gaussian pulse with a phase of 180°, 100 trajectories. cDissociation timeis when the distance of atoms in fragments reaches 5.3 Å.



dications dissociate during the laser pulse due to the light massof hydrogen, and thus we cannot completely rule out thepossibility of enhanced ionization. However, the laser pulseparameters can be optimized to retain high mode selectivitywhile reducing ionization, for example, with lower intensities.At the lower field strength of 0.07 au, the yield of H+ is 84%,while the yield of I+ is 9% (Table 2). Some of the remainingunfinished trajectories would eventually lead to energeticallyfavored I+ elimination via IVR. In this case, the averageddissociation time of H+ elimination is ∼114 fs, corresponding tothe second last laser cycle. Here we choose dissociation time aswhen the distance of atoms in fragments reaches 5.3 Å (10bohr). Because the laser intensities in remaining laser cycles arenot as high as the peak, charge-resonance-enhanced ionizationis less likely. We also varied the number of cycles and the phaseof a laser pulse to find optimum conditions. A four-cycle laserpulse with a phase of 180° achieves almost 60% yield, and thefast elongation of the C−H bond happens at the last laser cycle,which has relatively low field strengths. With this shorter pulse,dynamical ionization will be reduced significantly together withhydrogen ion yield. However, hydrogen ion production remainsthe dominant channel among all reaction pathways.In addition to optimizing the laser pulse for delayed H+

dissociation in C6H5I2+, other aspects of charge-resonance-

enhanced ionization should be taken into account. CREI istypically seen in the dissociation of diatomic molecules atwavelengths around 800 nm,24 and has been also demonstratedin the dissociation of CO2.

25 A similar mechanism wasproposed to explain enhanced ionization in polyatomichydrocarbon molecules at twice the equilibrium internucleardistance.26 Because of the much lower photon energy in themid-infrared range (∼0.12 eV for 10 μm), the criticalinternuclear distance for initiating CREI is near or in theasymptotic region. At these distances, the C−H bond is alreadybroken regardless of the potential ionization dynamics. A recentstudy of H2

+ dissociation by infrared pulses showed thationization is not a prevailing process (less than one-third of thedissociation probability), even though CREI still exists.21

Therefore, we can reasonably expect that field-induceddissociation of iodobenzene dication will occur in the mid-IRregion without substantial ionization.Electronic excitation during the laser pulse can also compete

with dissociation. In many case, this happens even beforesignificant ionization. Substantial nonresonant excitation fromthe ground state was found in TD-CIS simulations of butadienein strong fields of 800 nm laser pulses.27 In our previous studyof ClCHO+, the excitation probability is found to be negligibledue to the delayed dissociation and the low energy of the longwavelength photons.17 The delayed dissociation (field-freedissociation) prevents the excitation rate from increasingbecause the field is already off when the bonds are significantlyelongated and the energy levels are close to each other. Giventhis rationale, most of the CF3Br

+ trajectories dissociate withoutsignificant excitation, while the C6H5I

2+ trajectories would haveto be examined more closely because C6H5I

2+ has some low-lying excited states and most hydrogen ions dissociate withinthe pulse. The TD-CIS approach was employed with the 6-311G(d,p) basis set to calculate the population of the groundand excited electronic states at the equilibrium geometry. Toevaluate the excitation probabilities with elongated C−Hbonds, we sampled averaged geometries (e.g., C−H bondlength of ∼3 Å) from trajectories that correspond to the secondlast (or last) laser cycle and carried out TD-CIS calculations

with the remainder of the pulse. It turns out that the effects ofdeformed geometries on the excitation probability are small (ornegligible) due to low field strengths in remaining laser cycles.The calculated excitation probability is strongly dependent onthe field direction due to a large permanent dipole along theC−I bond. Considerable excitation occurs even at longwavelengths with the laser polarization parallel to the C−Ibond, while little or no excitation is seen with the polarizationperpendicular to the C−I bond for any of the field strengthsemployed in the study. For the four-cycle laser pulse with a fieldstrength of 0.07 au, 40% of the population of the ground state istransferred to excited states when the polarization is along theC−I bond. Of the remaining 60% in the ground state, morethan half should yield para-hydrogen elimination (Table 2).Dissociation will occur on the excited-state surfaces and couldyield hydrogen, but this is outside the scope of the presentstudy. Because no excitation is seen when the polarization isperpendicular to the C−I bond, this field orientation could beused to achieve ortho- or meta-hydrogen ion dissociationwithout excitation. Substantial H+ elimination can be achievedfor polarization perpendicular to the C−I bond if the higherfield strengths are used, for example, 0.09 au. For this case, thedegree of excitation is still small (Pe ≈ 11%) even afterconsidering the effect of elongated bonds. The averagedissociation time is at the second last cycle of the pulse, sothe ionization probability should also be small.

4. DISCUSSIONThe bond-selective mid-IR dissociation approach presentedhere exploits both the charge oscillation of molecules in stronglaser fields and large ponderomotive potentials afforded by thelong wavelength of a mid-IR laser. Significant chargeoscillations are present in molecular systems with any drivinglaser wavelength at high intensity because it is a pure electroniceffect due to the distortion of the electron cloud. However, it isthe ponderomotive potential that enables the energy transferfrom the laser directly to the kinetic energy of the nuclei. Withthe most widely used 800 nm lasers, while the charge oscillationcould be very large, the nuclei cannot follow the quickoscillation of the laser field and the electron density, and thusthe kinetic energy gain is indirect and inefficient, as reflected bya small ponderomotive potential. Therefore, it is not surprisingfrom this point of view that 800 nm Ti:sapphire lasers are notefficient in dissociating molecules without invoking ionizationor even Coulomb explosion. Only weak perturbation of thenuclear configuration is possible, as demonstrated withimpulsive Raman scattering excitation of vibrations. Coherentcontrol exploits both electronic and nuclear degrees of freedomwith intense Ti:sapphire lasers.7,8 While coherent controlcoupled to shaped laser pulses and genetic algorithm enjoyedgreat success in recent years, its underlying mechanisms areoften complicated due to the intertwined processes ofionization and dissociation.The concept of ponderomotive potential has been widely

used in the strong field physics community for calculating thekinetic energy of ionized electrons and the energy cutoff of highharmonic generation. The ponderomotive potential provides aconvenient indicator of how much kinetic energy a constituentatom can gain from a laser field, but it is not an upper bound tothe instantaneous kinetic energy gain because it is defined asthe cycle-averaged kinetic energy for a free charge. In principle,atoms in a molecule can gain as much as four times theponderomotive potential during only the first half of a laser



cycle before being slowed down during the second half cycle.As discussed in our previous paper, charge oscillation can allowthe constituent atoms to continue accelerating rather thanslowing down. The final energy gain, however, depends on theactual value of the charge (generally smaller than unit charge)and the details of dynamics (e.g., the phase of laser fields andwhether and when the nuclei rescatter from each other).Finally, it is worth pointing out the technical advancement of

producing intense and ultrashort mid-IR pluses required toimplement this new approach has been pursued intensively inthe ultrafast laser community for a different reason: using amid-IR laser to drive high harmonic generation to produceintense attosecond pulses. The approaches being investigatedinclude a free electron laser and optical parametrical chirpedpulse amplification (OPCPA).28,29 A more conventionalmethod to produce ultrashort mid-IR laser pulses is to use anoptical parametrical amplifier (OPA) pumped by a Ti:sapphirelaser. This technique has been widely used to produce milli-Joule level ultrashort pulses in the mid-IR range.30 Therefore,the implementation of the current approach can be expected inthe near future.

5. CONCLUSIONS

A novel mechanism for achieving the bond-selective dissocia-tion with a strong mid-IR laser pulse has been investigated inlarge polyatomic molecules (CF3Br

+ and C6H5I2+) using ab

initio classical trajectory calculations. We observed very largeenhancements of energetically disfavored reaction channels forboth systems. This was accomplished by choosing the laserpolarization to be parallel to the specific bond being cleaved.The wavelength dependence of the F yield in CF3Br

+

dissociation shows that resonant vibrational excitation doesplay a role in the dynamics of strong field dissociation. Theselective bond dissociation in C6H5I

2+ provides a goodcandidate for experimental implementation because a nearlyperfect alignment of this molecule has been realizedexperimentally. Favorable branching ratios can be obtained bychanging the driving laser polarization: pure para-hydrogenelimination occurs when the polarization is along the C−Ibond. The rapid development of laser technology in the mid-IRrange28−30 and intensive studies of molecular alignment31 andorientation23,32 of polyatomics should enable us to implementbond-selective dissociation experimentally using the proposedmethod, which is simple and general in controlling photo-dissociation of cations.

■ AUTHOR INFORMATION

Corresponding Author*E-mail: [email protected]. Tel: 1-313-577-8658

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

This work was supported by a grant from the National ScienceFoundation (CHE1212281) and Army Research Office. Wethank Wayne State University’s computing grid for computertime.

■ REFERENCES(1) Crim, F. F. State- and Bond-Selected Unimolecular Reactions.Science 1990, 249, 1387−1392.

(2) Nesbitt, D. J.; Field, R. W. Vibrational Energy Flow in HighlyExcited Molecules: Role of Intramolecular Vibrational Redistribution.J. Phys. Chem. 1996, 100, 12735−12756.(3) Schulz, P. A.; Sudbo, A. S.; Krajnovich, D. J.; Kwok, H. S.; Shen,Y. R.; Lee, Y. T. Multi-Photon Dissociation of Polyatomic-Molecules.Annu. Rev. Phys. Chem. 1979, 30, 379−409.(4) Bloembergen, N.; Zewail, A. H. Energy Redistribution in IsolatedMolecules and the Question of Mode-Selective Laser ChemistryRevisited. New Experiments on the Dynamics of Collisionless EnergyRedistribution in Molecules Possibilities for Laser-Selective Chemistrywith Subpicosecond Pulses. J. Phys. Chem. 1984, 88, 5459−5465.(5) Diau, E. W. G.; Herek, J. L.; Kim, Z. H.; Zewail, A. H.Femtosecond Activation of Reactions and the Concept of NonergodicMolecules. Science 1998, 279, 847−851.(6) Windhorn, L.; Yeston, J. S.; Witte, T.; Fuss, W.; Motzkus, M.;Proch, D.; Kompa, K. L.; Moore, C. B. Getting Ahead of Ivr: ADemonstration of Mid-Infrared Induced Molecular Dissociation on aSub-Statistical Time Scale. J. Chem. Phys. 2003, 119, 641−645.(7) Brumer, P.; Shapiro, M. Coherence Chemistry: ControllingChemical Reactions with Lasers. Acc. Chem. Res. 1989, 22, 407−413.(8) Assion, A.; Baumert, T.; Bergt, M.; Brixner, T.; Kiefer, B.;Seyfried, V.; Strehle, M.; Gerber, G. Control of Chemical Reactions byFeedback-Optimized Phase-Shaped Femtosecond Laser Pulses. Science1998, 282, 919−922.(9) Levis, R. J.; Menkir, G. M.; Rabitz, H. Selective Bond Dissociationand Rearrangement with Optimally Tailored, Strong-Field LaserPulses. Science 2001, 292, 709−713.(10) Nuernberger, P.; Wolpert, D.; Weiss, H.; Gerber, G.Femtosecond Quantum Control of Molecular Bond Formation. Proc.Natl. Acad. Sci. U. S. A. 2010, 107, 10366−10370.(11) Chelkowski, S.; Gibson, G. N. Adiabatic Climbing of VibrationalLadders Using Raman Transitions with a Chirped Pump Laser. Phys.Rev. A 1995, 52, R3417−R3420.(12) Bergmann, K.; Theuer, H.; Shore, B. W. Coherent PopulationTransfer among Quantum States of Atoms and Molecules. Rev. Mod.Phys. 1998, 70, 1003−1025.(13) Amstrup, B.; Henriksen, N. E. Control of Hod Photo-dissociation Dynamics Via Bond-Selective Infrared MultiphotonExcitation and a Femtosecond Ultraviolet-Laser Pulse. J. Chem. Phys.1992, 97, 8285−8295.(14) Elghobashi, N.; Gonzalez, L.; Manz, J. Quantum ModelSimulations of Symmetry Breaking and Control of Bond SelectiveDissociation of FHF− Using IR+UV Laser Pulses. J. Chem. Phys. 2004,120, 8002−8014.(15) Hause, M. L.; Yoon, Y. H.; Crim, F. F. Vibrationally MediatedPhotodissociation of Ammonia: The Influence of N–H StretchingVibrations on Passage through Conical Intersections. J. Chem. Phys.2006, 125, 174309/1−7.(16) Crim, F. F. Vibrationally Mediated Photodissociation -Exploring Excited-State Surfaces and Controlling DecompositionPathways. Annu. Rev. Phys. Chem. 1993, 44, 397−428.(17) Lee, S. K.; Suits, A. G.; Schlegel, H. B.; Li, W. A ReactionAccelerator: Mid-Infrared Strong Field Dissociation Yields Mode-Selective Chemistry. J. Phys. Chem. Lett. 2012, 3, 2541−2547.(18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,B.; Petersson, G. A.,et al. Gaussian Development Version Revision H.11;Gaussian, Inc.: Wallingford, CT, 2010.(19) Lee, S. K.; Li, W.; Schlegel, H. B. HCO+ Dissociation in a StrongLaser Field: An Ab Initio Classical Trajectory Study. Chem. Phys. Lett.2012, 536, 14−18.(20) Yudin, G. L.; Ivanov, M. Y. Nonadiabatic Tunnel Ionization:Looking inside a Laser Cycle. Phys. Rev. A 2001, 64, 013409/1−4.(21) Picon, A.; Jaron -Becker, A.; Becker, A. Enhancement ofVibrational Excitation and Dissociation of H2

+ in Infrared LaserPulses. Phys. Rev. Lett. 2012, 109, 163002/1−5.(22) Kumarappan, V.; Bisgaard, C. Z.; Viftrup, S. S.; Holmegaard, L.;Stapelfeldt, H. Role of Rotational Temperature in Adiabatic MolecularAlignment. J. Chem. Phys. 2006, 125, 194309/1−7.



mailto:[email protected]

(23) Holmegaard, L.; Nielsen, J. H.; Nevo, I.; Stapelfeldt, H.;Filsinger, F.; Kupper, J.; Meijer, G. Laser-Induced Alignment andOrientation of Quantum-State-Selected Large Molecules. Phys. Rev.Lett. 2009, 102, 023001/1−4.(24) Zuo, T.; Bandrauk, A. D. Charge-Resonance-EnhancedIonization of Diatomic Molecular-Ions by Intense Lasers. Phys. Rev.A 1995, 52, R2511−R2514.(25) Bocharova, I.; Karimi, R.; Penka, E. F.; Brichta, J.-P.; Lassonde,P.; Fu, X.; Kieffer, J.-C.; Bandrauk, A. D.; Litvinyuk, I.; Sanderson, J.;et al. Charge Resonance Enhanced Ionization of CO2 Probed by LaserCoulomb Explosion Imaging. Phys. Rev. Lett. 2011, 107, 063201/1−5.(26) Lotstedt, E.; Kato, T.; Yamanouchi, K. Efficient Ionization ofOne-Dimensional Acetylene Investigated by Time-Dependent Har-tree-Fock Calculations. Phys. Rev. A 2012, 86, 023401.(27) Sonk, J. A.; Caricato, M.; Schlegel, H. B. Td-Ci Simulation ofthe Electronic Optical Response of Molecules in Intense Fields:Comparison of RPA, CIS, CIS(D), and EOM-CCSD. J. Phys. Chem. A2011, 115, 4678−4690.(28) Ross, I. N.; Collier, J. L.; Matousek, P.; Danson, C. N.; Neely,D.; Allott, R. M.; Pepler, D. A.; Hernandez-Gomez, C.; Osvay, K.Generation of Terawatt Pulses by Use of Optical Parametric ChirpedPulse Amplification. Appl. Opt. 2000, 39, 2422−2427.(29) Andriukaitis, G.; Balciunas, T.; Alisauskas, S.; Pugzlys, A.;Baltuska, A.; Popmintchev, T.; Chen, M.-C.; Murnane, M. M.;Kapteyn, H. C. 90 Gw Peak Power Few-Cycle Mid-Infrared Pulsesfrom an Optical Parametric Amplifier. Opt. Lett. 2011, 36, 2755−2757.(30) Popmintchev, T.; Chen, M.-C.; Popmintchev, D.; Arpin, P.;Brown, S.; Alisauskas, S.; Andriukaitis, G.; Balciunas, T.; Mucke, O. D.;Pugzlys, A.; et al. Bright Coherent Ultrahigh Harmonics in the KeV X-Ray Regime from Mid-Infrared Femtosecond Lasers. Science 2012,336, 1287−1291.(31) Lee, K. F.; Villeneuve, D. M.; Corkum, P. B.; Stolow, A.;Underwood, J. G. Field-Free Three-Dimensional Alignment ofPolyatomic Molecules. Phys. Rev. Lett. 2006, 97, 173001/1−4.(32) Fleischer, S.; Zhou, Y.; Field, R. W.; Nelson, K. A. MolecularOrientation and Alignment by Intense Single-Cycle THz Pulses. Phys.Rev. Lett. 2011, 107, 163603/1−5.



bond-selective dissociation of polyatomic cations in mid-infrared strong fields · 2014-01-23 ·...

Documents